Ezzine L.,Electronic and Instrumentation Team |
Guerbaoui M.,Electronic and Instrumentation Team |
El Afou Y.,Electronic and Instrumentation Team |
Ed-dahhak A.,Electronic and Instrumentation Team |
And 3 more authors.
International Conference on Integrated Modeling and Analysis in Applied Control and Automation | Year: 2010
The goal of this paper is to carry out a statistical study whose objective is the identification of time series of the greenhouse climatic parameters, in order to optimize the expenditure in cost and time of the culture under greenhouse. In this study, we showed that the inside temperature is the most influential parameters on the greenhouse. However, the automatic climate control requires the development of appropriate control laws that are based on models representing linear and nonlinear system. We are therefore forced to make a study of the system to generate a model that faithfully reproduces the operating parameters of greenhouse climate. In order to achieve the maximum benefit it is important to exploit the available data and an obvious choice here are the machine learning methods such as artificial neural networks. The use of recurrent Radial Basis Function (RBF) models is justified by employing a nonlinear greenhouse system, and hence to give the possibility to identify and to control in the real time the inside temperature in the greenhouse, taking into accounts other climatic parameters within and outside the greenhouse. A comparison of the measured and simulated data proved that the found model can envisage correctly the inside greenhouse temperature.